numerical-robustness

Detect and resolve floating-point issues in CAD kernel operations.

20|3|Updated Mar 2, 2026

Search Tags:#CAD #floating-point #boolean operations #numerical robustness #CAD kernel

Ready to run

npx skills add https://github.com/andymai/brepkit --skill numerical-robustness

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: numerical-robustness
Source: https://github.com/andymai/brepkit/tree/main/.claude/skills/numerical-robustness
Command: npx skills add https://github.com/andymai/brepkit --skill numerical-robustness

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps debug and prevent floating-point issues in the CAD kernel, ensuring stability and accuracy in geometric calculations.

Core Features & Use Cases

Floating-Point Failure Detection: Identifies issues like inconsistent boolean operations, collapsed AABBs, and near-duplicate vertices.
Symptom Analysis: Provides a symptom-to-cause table for common issues, such as pass/fail variations and near-duplicates.
Rules for Robustness: Offers a set of rules to follow for numerical stability in CAD operations.
Quick Procedures: Includes commands for confirming cross-process nondeterminism and probing bucket-straddle issues.
Anti-Patterns: Advises against common pitfalls in handling floating-point comparisons and geometric computations.
Related Skills: Cites related skills for further debugging and verification.

Quick Start

Run the numerical-robustness skill to analyze and resolve numerical issues in your CAD kernel operations.

Frequently Asked Questions about numerical-robustness

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I fix inconsistent boolean operations in my CAD kernel?▼

Inconsistent boolean operations in your CAD kernel are typically caused by floating-point issues. You can resolve them by applying numerical robustness checks to detect near-duplicate vertices and prevent geometric calculation errors.

Why does my CAD geometry fail nondeterministically across processes?▼

Cross-process nondeterminism in CAD geometry usually stems from numerical instability. You can confirm this issue using specific probing commands to identify bucket-straddle problems and floating-point comparison failures.

What causes collapsed AABBs during complex geometric calculations?▼

Collapsed AABBs during geometric calculations occur due to lost floating-point precision. Implementing strict numerical robustness rules ensures geometric accuracy and prevents bounding volume collapses in your CAD workflow.

What are common anti-patterns for floating-point comparisons in CAD?▼

Common anti-patterns for floating-point comparisons in CAD include direct equality checks and naive geometric computations. Avoiding these pitfalls ensures numerical stability and prevents near-duplicate vertex issues.

When do I need numerical robustness checks for complex CAD geometry?▼

You need numerical robustness checks for complex CAD geometry when performing boolean operations that are sensitive to floating-point precision. This ensures stability and accuracy in workflows involving near-duplicate vertices and bounding volumes.